The Admin UI clearly shows the Sentiment Configuration updating correctly, yet the Go client returns a 400 Bad Request when submitting the query payload to /api/v2/analytics/conversations/queries. I’ve aligned the Interaction ID list with the Data Retention Policy settings, but the model availability flag still triggers a validation mismatch. Cursor-based pagination handles smaller batches without issue, though the normalization logic degrades once the score scaling algorithm processes raw polarity shifts. We don’t receive the expected webhook callbacks to the coaching platform, which renders the audit log generation functionally useless. The Queue Analytics dashboard remains idle during the aggregation phase because the Throttle Configuration blocks the batch size.
Here is the request struct handling the cursor navigation. The normalization step divides the raw score by the benchmark threshold, but the response payload drops the sentimentGranularity field entirely. It’s pretty baffling how the audit log generation stalls here.
type SentimentQuery struct {
InteractionIDs []string `json:"interactionIds"`
TimeRange string `json:"timeRange"`
Granularity string `json:"sentimentGranularity"`
}
The webhook payload validation fails at the third retry.